GBM Model Reproducibility

subbiah_an Member Posts: 2

I am building GBM regression model. I am explicitly setting the seed in both places - data split as well as model config. Still, my predictions are different every time I run the model. How to make my model reproducible? I am running this model in Databricks and the cluster/ core/ node set up all remain the same. Validation or hold out dataset also remain the same. Still model not producible?